As of 1960-64 the remaining twelve countries of pretransition status were moving toward the threshold ranges, but they were not yet (in 1960-64) at a stage where early fertility reduction would be predicted. The applicability of the socioeconomic threshold as a predictive device needs more testing, including longitudinal analysis, but the results so far suggest considerable stability in the relation between socioeconomic development and the onset of fertility declines. A striking feature of the Latin American situation is the high levels of development required for natality transition. Major fertility declines occurred in Europe at much lower levels of literacy, life expectancy, and proportion of the population employed outside agriculture (21). A possible explanation lies in the fact that progress along these lines in Europe was the result of a long and slow historical process, by modem standards. In Latin America many aspects of socioeconomic modernization are occurring much more rapidly than they did in Europe, with less time to change attitudes and practices affecting family size. Urbanization, in statistical terms, is occurring far more rapidly than it did in Europe. By the same token it is reasonable to expect more rapid reduction of fertility, once begun, because of modem technologies of communication and of family limitation, and, indeed, this seems to be happening. The threshold hypothesis has not yet been examined in detail for other regions but some preliminary conclusions follow. East and Southeast Asia In east and southeast Asia* there are also significant relationships between socioeconomic indicators and the birth rate. As is illustrated in Table 6, there appears to be a closer negative relationship between measures of economic achievement and birth rates than in Latin America. This is suggested by the higher correlation coefficients for per capita income and energy consumption in east and southeast Asia. As might be expected, the threshold of socioeconomic development accompanying fertility decline in this region is substantially lower than the threshold in Latin America. Of the seven areas clearly in natality transition by 1960-641 all had experienced the onset of fertility decline with per capita incomes well below $300, roughly the average level for onset of fertility ^Including twenty-three areas bounded by Pakistan, Indonesia, and Japan and excluding Asian parts of the Soviet Union. Usable data on most variables are available for only seventeen. All the coeffieicnts presented are statistically significant at the 5 percent level and most at the 1 percent level or better. Nevertheless the countries included arc few and scarcely a random sample, so weight should be given to general levels and patterns, not to specific values. ti------ *!»„ t>.,,.i,.,..„ Tn;..,n~ u~.,,, !/•„„„ \/i*in,,r, o:-----«,.,„ .....i r>~,,i.*..atly varying quality, single figures should not be taken very seriously. Thus the extraordinarily high correlation of telephones (i.e., a surrogate for modern infrastructure) with natality is reduced from -0.94 to -0.88 when Argentina and Uruguay arc excluded.